Which healthcare AI startups are making the most revenue today?
In our healthcare AI market deck, you will find everything you need to understand the market
SUMMARY
OpenEvidence appears to be making the most revenue among private healthcare AI startups today, running close to $300 million on an annualized basis, with Sword Health and Innovaccer forming the next tier.
The leaderboard is already big enough to matter: several private healthcare AI companies are above $100 million in recurring or annualized revenue, and a few have pushed past $200 million.
The biggest surprise is that ambient scribes do not dominate the revenue ranking. Clinical search and pharmaceutical advertising, AI-enabled care, healthcare data infrastructure and payer automation are producing the largest revenue pools.
OpenEvidence has the fastest-looking monetization engine because physicians can adopt it without a hospital buying it first, while pharmaceutical advertisers pay for access to that verified clinical audience.
Sword Health shows a very different path to scale. It makes more money per user by helping deliver care itself, but that comes with a more operationally complicated model than software or advertising.
Innovaccer's position is easy to underestimate because it predates the generative-AI boom. Its advantage is distribution: once a company already sits inside health-system data and workflows, new AI products can be sold on top of connections that are already there.
Revenue quality varies almost as much as revenue size. ARR, contracted ARR, annualized monthly revenue and recognized revenue are all useful, but they are not interchangeable, so the ranking is strongest at the tier level rather than at fake single-digit precision.
Abridge is the clearest commercial leader in ambient clinical documentation, yet the latest hard evidence still puts it around the $100 million-plus level rather than proving a $200 million current business.
Usage alone can be misleading. Aidoc, Viz.ai, Heidi and Nabla reach enormous numbers of hospitals, clinicians or patients, but that does not automatically translate into the same revenue scale as companies attached to larger existing budgets.
The most durable revenue may sit with companies embedded in operational workflows such as data infrastructure, prior authorization, patient communications and clinical documentation. OpenEvidence has a different moat: physician habit, audience scale and unusually attractive advertising economics.
The broader pattern is that healthcare AI companies are reaching nine-figure revenue much faster than earlier healthcare software businesses did. The winners are not creating entirely new budgets; they are attaching AI to spending categories that were already enormous.
Why is healthcare AI revenue worth ranking now?
Healthcare AI has finally produced several private companies making well over $100 million a year, so today we are comparing real businesses rather than promising pilots.
Bessemer Venture Partners' State of Health AI 2026 highlighted how unusually quickly some healthcare AI companies are reaching $100 million or even $200 million of annual recurring revenue. Earlier generations of healthcare software often needed a decade or more to reach those levels. Some of the new AI companies are getting there within five years.
The interesting part is where the revenue is coming from. OpenEvidence makes money from pharmaceutical advertising around a clinical-information product. Sword Health sells AI-enabled care to employers and health plans. Innovaccer sells healthcare data infrastructure and AI workflows. Abridge sells ambient clinical documentation to health systems. Cohere Health automates work for insurers.
Healthcare AI is therefore large enough now that "who makes the most money?" has become a real ranking question. The answer also tells us which healthcare problems customers are already willing to spend heavily to automate.
What should count as a healthcare AI startup?
For this ranking, we count private healthcare companies where AI is central to the product or to how the company delivers its service.
That includes companies such as OpenEvidence, Sword Health, Innovaccer, Abridge, Cohere Health, Artera and Aidoc. We do not require every dollar of revenue to come from a generative-AI feature. Sword, for example, still involves human clinicians, but AI sits at the center of its care model and its Phoenix system increasingly handles parts of care autonomously.
At the other extreme, we would not include every old healthcare software company simply because it has added an AI assistant.
We also exclude public companies. Tempus AI shows why that boundary matters. Tempus generated $382.5 million of revenue in one recent quarter and lifted its annual guidance to around $1.6 billion. If we treated Tempus as a startup, the private-company comparison would become almost pointless.
This market map, featured in our healthcare AI market deck, highlights top companies and startups in the healthcare AI market
Can we really compare all these healthcare AI revenue numbers?
Only approximately, because healthcare AI companies routinely disclose ARR, contracted ARR and annualized revenue alongside normal revenue, and those figures measure different things.
ARR usually annualizes recurring business already being generated. Contracted ARR can include signed customers whose deployments have not fully started. A run rate may simply multiply the latest month by twelve. Actual revenue records what the company recognized over a completed period.
The distinction is most important for companies growing extremely quickly. OpenEvidence's current figure comes from annualizing recent monthly advertising sales. Innovaccer reports ARR. Artera reports contracted ARR. Abridge has disclosed both ARR and contracted ARR through reporting around its financing.
So we can rank broad commercial scale with reasonable confidence, but pretending that Company A is exactly $8 million ahead of Company B would be fake precision.
| Company | Latest useful revenue measure | What we actually know |
|---|---|---|
| OpenEvidence | Close to $300M | Annualized revenue |
| Sword Health | ~$240M+ | Last clearly reported annual run rate |
| Innovaccer | >$200M | ARR |
| EliseAI | $200M | Company-wide ARR across housing and healthcare |
| Abridge | ~$100M-$117M | ARR / contracted ARR |
| Artera | >$100M | Contracted ARR |
| Aidoc | ~$60M | Estimated annual revenue |
Which healthcare AI startup is making the most revenue today?
OpenEvidence is currently the strongest candidate for the highest-revenue private healthcare AI startup, generating close to $300 million on an annualized basis.
Recent reporting from The Information put OpenEvidence at roughly $25 million of monthly revenue. Seven months earlier, the company was doing about half that rate.
The speed of the increase is what separates OpenEvidence from most of the field. Sacra estimates that annualized revenue was only $7.9 million in 2024, while Forbes was reporting roughly $50 million by mid-2025. It then climbed toward $150 million around the end of the year before roughly doubling again.
That progression puts OpenEvidence ahead of the latest disclosed numbers we can verify for Sword Health and Innovaccer.
There is one caveat worth keeping: annualized revenue is not trailing twelve-month revenue. But if the question is who appears to be making the most money right now, the latest monthly pace is the best evidence we have.
If you want more recent data on this point, please see our latest healthcare AI market report.
As this chart shows, and as featured in our healthcare AI market deck, search interest in healthcare AI has grown rapidly
Why is OpenEvidence growing so fast?
OpenEvidence is growing unusually fast because doctors can start using the product without waiting for their hospital to buy it, while pharmaceutical companies pay to reach those doctors.
That removes a major bottleneck in healthcare software.
OpenEvidence says more than 860,000 licensed U.S. clinicians now use the platform. Clinical consultations increased from roughly 3 million a month to about 18 million in December 2025 and then reached approximately 20 million in January 2026. In March, clinicians conducted one million consultations on a single day.
That audience is exceptionally valuable to pharmaceutical advertisers because these are verified medical professionals searching for clinical information. OpenEvidence can therefore let doctors use the product free while charging drug companies for advertising around those searches.
Compare that with enterprise healthcare AI. Abridge may need security reviews, EHR integration, contracting and a system-wide deployment before a large hospital starts paying. Another doctor can adopt OpenEvidence independently.
That difference helps explain why OpenEvidence has produced a revenue curve that looks much closer to a fast internet platform than a normal healthcare SaaS company.
Is OpenEvidence's revenue actually as attractive as the headline suggests?
OpenEvidence currently has some of the best economics we have seen in healthcare AI, with reported gross margins around 90% and the company operating near cash-flow breakeven.
The Information also reported that OpenEvidence was selling advertising against less than 5% of its available inventory. That is a striking number. It suggests the company reached its present scale without filling the product with ads.
There are limits to how far we can stretch that observation. Selling five or ten times more inventory could lower advertising prices or make physicians dislike the product. Pharmaceutical advertising budgets are large but finite.
Still, the combination is unusual: a free product with massive professional usage, premium advertising buyers and roughly software-level margins. Many healthcare AI companies would need substantially more revenue to produce the same gross profit dollars.
This chart, featured in our healthcare AI market deck, shows annual VC investment in healthcare AI startups
Could Sword Health actually be the biggest healthcare AI company?
Sword Health is probably the largest private healthcare AI company focused on delivering care, with its last clearly reported annual revenue run rate at roughly $240 million before its business expanded further.
TechCrunch reported that figure when Sword raised capital at a $4 billion valuation in 2025. Unlike OpenEvidence, Sword earns money by providing AI-assisted healthcare programs to employers and health plans, particularly in musculoskeletal care.
The company has kept expanding. Sword acquired Kaia Health for $285 million in early 2026, adding a large digital musculoskeletal business and giving Sword a stronger presence in Europe. It has also moved beyond its original physical-therapy positioning into pelvic health, mental health and broader AI Care.
Sword says its platform has delivered more than 10 million AI Care sessions and reached more than 700,000 members.
The old $240 million run-rate figure is therefore likely below Sword's current scale, but we do not have a newer disclosed revenue number strong enough to replace it. For now, Sword belongs securely in the top group without pretending we know whether the business is at $260 million, $300 million or something else.
If you want more recent data on this point, please see our latest healthcare AI market report.
How did Innovaccer quietly get above $200 million in ARR?
Innovaccer has now crossed $200 million in ARR, making the healthcare data company one of the largest private AI businesses in the sector.
Fortune recently reported that Innovaccer increased ARR from roughly $130 million to more than $200 million. That works out to growth of more than 50%, even from an already substantial base.
Innovaccer looks different from the newer generative-AI companies because it spent years building healthcare data infrastructure before the current AI boom. Its platform pulls information from electronic health records, claims and other systems, then makes that data usable for care management, operations and AI applications.
The scale underneath that business is considerable. Innovaccer works with seven of the ten largest U.S. health systems and says its platform touches around 80 million patient records.
That existing infrastructure gives Innovaccer something many young healthcare AI startups still have to build: distribution inside large health systems. New AI products can increasingly be sold on top of relationships and data connections that are already there.
This chart, featured in our healthcare AI market deck, looks at Tempus AI’s strategy in healthcare AI
Does EliseAI really belong among the biggest healthcare AI startups?
EliseAI has reached $200 million ARR, but we cannot call it a $200 million healthcare AI business because a substantial part of that revenue still comes from housing.
EliseAI announced in June 2026 that company-wide ARR had doubled to $200 million, marking its fifth consecutive year of roughly 100% annual growth. The company began by automating leasing, maintenance and resident communications for property managers before taking the same conversational AI approach into healthcare.
Its healthcare product now handles tasks such as scheduling, intake and patient communications.
The problem is simply that EliseAI does not disclose how much of its $200 million comes from healthcare. Assigning the full figure to this ranking would overstate its position.
We should still keep EliseAI in the conversation because $200 million of company-wide ARR proves that its underlying AI workflow platform has reached serious scale. But without segment revenue, its precise healthcare ranking remains unknowable.
Is Abridge already making $200 million a year?
Abridge is the biggest dedicated ambient clinical-AI startup we can measure, but the public evidence still does not prove that it has reached $200 million of current revenue.
The clearest reported figures put Abridge at roughly $100 million of ARR and around $117 million of contracted ARR during 2025. Those numbers already represent extremely fast growth: earlier reporting had the company at around $50 million.
Abridge has meanwhile become deeply embedded in major health systems. Kaiser Permanente announced deployment across roughly 24,600 physicians, while Abridge has also won organizations including Mayo Clinic, UPMC, Johns Hopkins and Duke.
The Information reported that management had presented investors with a plan to reach roughly $200 million of contracted ARR and later about $400 million. Those targets make the trajectory easy to understand, but targets are not revenue.
Until Abridge gives us a fresher number, the cleaner conclusion is that it is already a $100 million-plus business and probably the most commercially advanced pure ambient-AI company.
If you want more recent data on this point, please see our latest healthcare AI market report.
This chart, featured in our healthcare AI market deck, shows annual funding in healthcare AI startups
Are Artera and Cohere Health already bigger than most healthcare AI startups?
Artera and Cohere Health both appear to have crossed the $100 million scale, even though neither gets as much attention as the hottest clinical AI startups.
Artera said it had surpassed $100 million of contracted annual recurring revenue while serving more than 1,000 healthcare organizations. Its platform supports billions of patient-provider communications each year, and the company has been turning more of those conversations into automated AI workflows.
Cohere Health is harder to pin down precisely. CB Insights placed its revenue around $100 million in 2024, while Cohere subsequently said committed ARR grew more than 60% that year. The company processes roughly 12 million prior-authorization requests annually and serves more than 16 million health-plan members.
Cohere has also expanded beyond outpatient prior authorization into inpatient utilization management, payment integrity and policy management. That gives it several ways to increase spending within the same payer customers.
So neither company should be casually ranked at an invented $150 million or $180 million. What the evidence does support is placing both among the relatively small group of private healthcare AI businesses already operating around or above nine figures of recurring revenue.
Why do Aidoc and Viz.ai make less money despite reaching so many hospitals?
Aidoc and Viz.ai show how a healthcare AI company can achieve huge clinical distribution without automatically generating hundreds of millions of dollars in revenue.
CB Insights estimated Aidoc at roughly $60 million of annual revenue in 2025. Yet Aidoc says its technology is deployed in more than 1,600 medical centers and analyzes around 60 million patients each year. The company also raised another $150 million in 2026 to expand its CARE foundation model and aiOS platform.
Viz.ai operates at a similarly impressive clinical scale. The company has approached 2,000 hospitals, says its network covers more than 230 million lives and recently reported that its healthcare business had become profitable. Its life-sciences business has also grown quickly.
Those usage figures sound enormous beside a company generating $60 million or $100 million of revenue. The reason is that diagnostic and clinical-coordination platforms do not necessarily capture much money each time their software touches a patient.
A product can become deeply embedded in medicine before becoming one of the sector's largest revenue businesses.
This chart, featured in our healthcare AI market deck, compares the main business model options for ambient AI companies
Who is winning the AI medical-scribe revenue race?
Abridge is currently the clear commercial leader among independent ambient AI documentation startups, while Ambience, Heidi, Freed and Nabla remain considerably smaller on the revenue figures we can actually observe.
Abridge had already reached roughly nine figures of ARR on the latest solid disclosure. Sacra estimated Ambience Healthcare at about $30 million ARR in May 2025, up from roughly $19 million at the end of 2024. Ambience has kept winning large hospital deployments since then, so that figure is almost certainly stale, but the company has not published a cleaner replacement.
Freed was estimated around $19 million ARR in early 2025. Heidi Health has been estimated in the low-$20-million range despite reporting more than two million patient visits a week. Nabla has spread across tens of thousands of clinicians and more than 100 healthcare organizations without disclosing enough financial information for a useful ranking.
The striking part is the gap between usage and monetization. Several ambient AI products have reached enormous clinician audiences, yet Abridge has so far converted large enterprise deployments into revenue much faster than the rest.
| Ambient AI company | Best useful revenue indication | Where it stands |
|---|---|---|
| Abridge | ~$100M+ | Clear commercial leader |
| Ambience | ~$30M+ estimated ARR | Fast-growing enterprise challenger |
| Heidi | Low-$20M range estimated | Huge usage, lighter monetization |
| Freed | ~$19M estimated ARR | Strong self-serve business |
| Nabla | Undisclosed | Large footprint, unclear revenue |
Which healthcare AI business model is making money fastest?
Clinical search is monetizing fastest right now, while AI-enabled care, data infrastructure and payer automation are producing some of the largest and probably stickiest contracts.
OpenEvidence has the simplest growth engine. Doctors use the product free, which lets adoption spread without hospital procurement. Pharmaceutical companies then pay for access to that audience.
Sword attacks a much larger pool of spending by helping deliver healthcare itself. That creates more revenue per member but also means running a more operationally complicated business.
Innovaccer and Cohere Health go after deeply embedded enterprise workflows. Their sales cycles can be slower, but a large health system or insurer may keep paying for years once the technology is connected to core data and operations.
Abridge sits somewhere between those models. It begins with an immediately understandable product — automatically documenting the medical visit — then tries to expand into coding, billing, clinical intelligence and other workflows around the same encounter.
The companies generating the most revenue have one thing in common: they have attached AI to budgets that already existed. Pharmaceutical advertising, healthcare delivery, hospital software, insurer administration and clinician labor were all enormous spending categories long before generative AI arrived.
If you want more recent data on this point, please see our latest healthcare AI market report.
This chart, featured in our healthcare AI market deck, illustrates how revenue is distributed across customer segments in the healthcare AI market
Which healthcare AI startups have the safest revenue?
Innovaccer, Cohere Health, Artera and Abridge probably have some of the hardest revenue to displace because customers integrate these products directly into everyday healthcare operations.
Innovaccer connects health-system data across EHRs, claims and other systems. Replacing that infrastructure can become a multi-year technical project.
Cohere sits inside prior authorization and payer decision-making. Those are regulated, high-volume processes that insurers cannot simply turn off while trying another vendor.
Artera already owns communication workflows across more than 1,000 healthcare organizations. Every additional AI agent can be sold into an existing communication layer rather than starting with a completely new deployment.
Abridge benefits from a similar effect at the clinician level. Once thousands of doctors use the same ambient system every day and the software is connected to Epic or another EHR, switching involves more than comparing which model writes a nicer note.
OpenEvidence's moat works differently. Its strength comes from physician habit and audience scale rather than complicated enterprise integration. That can produce spectacular margins, although user habits can theoretically change faster than a hospital can rip out core infrastructure.
Profitability also differs sharply. Recent reporting describes OpenEvidence as roughly cash-flow breakeven. Artera has said it is profitable, and Viz.ai has said its healthcare business has reached profitability. Many of the heavily funded clinical AI companies still disclose too little for us to judge their operating economics confidently.
Is healthcare AI revenue really growing faster than old healthcare software?
Yes. The best healthcare AI startups are reaching $100 million of recurring revenue much faster than traditional healthcare software companies usually did.
Bessemer's healthcare analysis found that some of the new companies are reaching $100 million or even $200 million ARR within five years, whereas earlier healthcare software businesses often took ten years or longer.
The company-level trajectories back that up. OpenEvidence went from single-digit millions of annualized revenue in 2024 to tens of millions and then hundreds of millions. Abridge roughly doubled from around $50 million to about $100 million ARR in a short period. Innovaccer moved from approximately $130 million to beyond $200 million ARR in roughly a year.
These companies are also growing through different mechanisms, which makes the pattern harder to dismiss as one temporary boom. One monetizes physician attention, another automates documentation, another sells healthcare data infrastructure.
AI has clearly shortened the time it takes the best healthcare technology companies to turn adoption into meaningful revenue.
This chart, featured in our healthcare AI market deck, shows how symptom checker app technology has evolved over time
Which healthcare AI startups could cross $200 million next?
Abridge is the clearest candidate to become the next confirmed $200 million healthcare AI business, while Cohere Health and Artera have enough installed scale to get there through expansion rather than starting from scratch.
Abridge has the most visible trajectory. Its ambient documentation product is already deployed across some of the largest U.S. health systems, and the company is pushing into coding, revenue cycle, clinical intelligence and life-sciences workflows. That gives it several ways to increase revenue from the same underlying clinician interactions.
Cohere Health has enormous transaction volume to monetize. Its platform already handles millions of prior authorizations each year, and the company is moving into additional payer workflows where customers spend heavily on manual labor.
Artera has another advantage: more than 1,000 existing healthcare organizations and billions of patient communications already running through its platform. A new agent can therefore become an upsell inside an account the company already owns.
The next company to cross $200 million may come from selling more automation into an existing workflow rather than finding millions of brand-new users.
So which healthcare AI startups are making the most revenue today?
OpenEvidence is the clearest revenue leader among private healthcare AI startups today, followed by a small group led by Sword Health and Innovaccer, while Abridge, Artera and Cohere Health form the next major tier.
As seen above, OpenEvidence is running close to $300 million of annualized revenue. Sword Health's latest hard number was around $240 million before the company expanded through Kaia Health, while Innovaccer has now moved beyond $200 million ARR after growing from roughly $130 million.
Behind them, the ranking becomes less exact. Abridge has the strongest disclosed revenue among pure ambient AI companies. Artera is above $100 million of contracted ARR. Cohere Health also appears firmly inside the nine-figure group, although its latest exact revenue is not public.
EliseAI deserves separate treatment because its company-wide ARR has reached $200 million but healthcare represents only part of that total. Aidoc, Ambience, Heidi, Freed, Nabla and several other well-known clinical AI companies either remain below the nine-figure level on available estimates or do not disclose enough financial data for a clean ranking.
The surprising result is how little the revenue leaderboard resembles the healthcare AI hype leaderboard. Ambient scribes dominate attention, yet the biggest private revenue pools today come from clinical search and advertising, AI-enabled care and healthcare data infrastructure.
That gives us a fairly clear answer for now. OpenEvidence appears to be making the most revenue, Sword Health and Innovaccer sit immediately behind it, and healthcare AI has already produced several genuinely large private businesses across completely different parts of the healthcare system.
If you want more recent data on this point, please see our latest healthcare AI market report.
In our healthcare AI market deck, we identify pain points entrepreneurs should prioritize
OUR METHODOLOGY
This ranking asks which private healthcare AI startups are making the most revenue today. We treated it as an evidence-ranking problem because private companies disclose unevenly, revenue measures differ, and market visibility often tracks hype more closely than commercial scale.
We first defined the category narrowly enough to stay useful: private healthcare companies where AI is central to the product or to how the service is delivered. We excluded public companies such as Tempus AI and did not automatically count every established healthcare software company that has added an AI feature.
For each company, we prioritized the freshest credible evidence on present commercial scale. Recognized revenue, ARR, contracted ARR and annualized monthly revenue are all included when useful, but the underlying measure stays visible because those figures are not interchangeable. Where a company operates across several industries, as EliseAI does, we did not assign all company-wide revenue to healthcare.
We also used deployment scale, customer base, workflow depth, acquisitions, business mix and recent growth to interpret the headline number. Those factors help separate large usage from actual monetization and help flag revenue figures that may already be stale in a fast-growing company.
Key sources include Bessemer Venture Partners' State of Health AI 2026, The Information on OpenEvidence's revenue and advertising economics, TechCrunch on Sword Health's revenue run rate, Fortune on Innovaccer crossing $200 million ARR, The Information on Abridge's ARR trajectory, TechCrunch on Abridge's contracted ARR, Artera on surpassing $100 million in contracted ARR, and Cohere Health on committed ARR growth and transaction scale.
We also used direct company material for operating scale and business context, including Sword Health, Sword's Kaia Health acquisition, EliseAI, Abridge, Aidoc, Viz.ai, Nabla, Heidi Health, and Tempus AI for the public-company comparison.
The final order comes from where the strongest recent evidence converges, not from one database or a single revenue convention. That is why the ranking makes a strong call at the top while keeping some of the middle positions deliberately approximate.
This chart, featured in our healthcare AI market deck, illustrates how revenue is distributed geographically across Europe, Asia, North America, Africa, and South America in the healthcare AI market
Related blog posts
